Coinbase outage halts trading 5+ hours

- Coinbase users lost access to trading on web, mobile, and Exchange late May 7 after an AWS outage hit core systems and forced markets offline. - The disruption lasted roughly 8.5 hours on Coinbase.com and about 6.7 hours on Coinbase Exchange, with trading restored after cancel-only and auction modes. - The outage exposed how even “always-on” crypto venues still share old centralized infrastructure risks — and why execution risk remains very real.

Crypto is supposed to trade 24/7. That is the whole pitch — no closing bell, no market holidays, no waiting for Monday. But late on May 7, Coinbase users ran into the oldest problem in finance anyway: the market was there, prices were moving, and they still could not trade. The immediate cause was not a hack or a crypto-specific failure. It was Amazon Web Services. A heat-related incident in AWS’s Northern Virginia footprint knocked into Coinbase’s systems hard enough that web, mobile, and exchange trading stopped working for hours. Coinbase kept repeating one thing through the outage — customer funds were safe. That mattered, but it did not solve the core problem: if you cannot transact during a volatile market, safety is only half the story. (status.coinbase.com) ### What actually broke? Coinbase first posted at 5:56 p.m. PDT on May 7 that customers might be unable to transact on web and mobile. A few minutes later, its Exchange status page said users could be unable to transact there too. By 6:53 p.m. PDT, Coinbase had tied the issue to an AWS outage, and by 7:37 p.m. PDT it said the disruption came from increased temperatures in the affected AWS service. (sta([status.coinbase.com) Why did AWS matter so much? AWS said one availability zone in US-EAST-1 — use1-az4 — was dealing with increased temperatures after a thermal event and power loss. That hit EC2 instances and EBS volumes in the affected hardware. Basically, this was a cloud plumbing problem in one of the internet’s busiest regions. If a trading platform depends on that plumbing at the wrong layer, the outage stops feeling “technical” and starts feeling like a market halt. (health.aws.amazon.com) ### How long was Coinbase down? For Coinbase.com and the iOS and Android apps, the outage window ran from 5:56 p.m. PDT on May 7 until 2:28 a.m. PDT on May 8, when Coinbase said all markets had been re-enabled and customers could log in to trade. That is about 8 hours and 32 minutes. Coinbase Exchange showed a slightly shorter but still huge disruption — from 6:06 p.m. PDT until 12:49 a.m. PDT, or about 6 hours and 43 minutes. (status.coinbase.com) ### Why the weird “cancel only” and “auction” phases? Because you do not just flip a market back on after a long outage. Coinbase first moved markets into “Cancel Only” mode, where resting orders could be canceled but new market and limit orders were not accepted. Then it shifted into auction mode, where customers could post limit orders and see an indicative open price, but no matches happened for at l(status.coinbase.com)eopening traffic lane by lane instead of dropping everyone onto the highway at once. (status.exchange.coinbase.com) ### Was this only a Coinbase problem? No. Reuters said CME also reported platform issues during the same AWS incident, though it was not immediately clear whether every issue traced back the same way. That matters because it turns this from “Coinbase had downtime” into “large financial platforms still share infrastructure choke points.” Crypto likes to market itself as decentralized, but the user experience often runs through very centralized cloud providers. (money.usnews.com) ### Why do traders care so much? Because the real loss in an outage is often optionality. You may not lose custody of funds, but you lose the ability to react. If prices swing while one venue is dark, users cannot enter, exit, hedge, or rebalance. And when trading comes back, the restart itself can be messy — pent-up orders, price gaps, and a rush to catch up. That is execution risk, plain and simple. (status.exchange.coinbase.com) ### Does this change anything bigger? It sharpens a point the crypto industry keeps running into: nonstop markets still depend on ordinary infrastructure, and ordinary infrastructure fails. Coinbase’s systems were supposed to be resilient to a single-zone problem, but the incident spilled across enough dependencies to keep core trading down for hours. That is not a philosophical issue. It is a design issue — and users notice it fastest when the market is moving. (health.aws.amazon.com) ### Bottom line? The headline is not just that Coinbase went down. It is that one cloud-region heat event was enough to sideline trading at one of crypto’s biggest venues for most of a night. In a market built on the promise of constant access, that is the kind of failure people remember.

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.